Update app.py

app.py

@@ -424,3 +424,46 @@ if uploaded_file is not None:
                             color_continuous_scale='RdBu')  # Use a valid Plotly colorscale
             fig.update_layout(title="Word Length Correlations by Position")
             st.plotly_chart(fig)
+
+    with st.expander("Folio Clustering Based on Word Usage Patterns"):
+        st.write("""
+        This section groups folios into clusters based on their word usage patterns.
+        - **PCA**: Reduces the data to 2D using Principal Component Analysis.
+        - **t-SNE**: Reduces the data to 2D using t-Distributed Stochastic Neighbor Embedding.
+        - **K-Means**: Groups folios into clusters based on their word frequencies.
+        """)
+        
+        # Feature Extraction
+        all_words = set(word for folio in folio_word_map for word in folio_word_map[folio])
+        word_freq_matrix = pd.DataFrame(index=folio_word_map.keys(), columns=all_words, data=0)
+
+        for folio, word_counter in folio_word_map.items():
+            for word, count in word_counter.items():
+                word_freq_matrix.loc[folio, word] = count
+        
+        # Dimensionality Reduction Option
+        dim_reduction_method = st.selectbox("Select Dimensionality Reduction Method", ["PCA", "t-SNE"], key="dim_reduction_method")
+        
+        if dim_reduction_method == "PCA":
+            reducer = PCA(n_components=2)
+            folio_coords = reducer.fit_transform(word_freq_matrix)
+        else:
+            reducer = TSNE(n_components=2, random_state=42)
+            folio_coords = reducer.fit_transform(word_freq_matrix)
+        
+        # Clustering (K-Means)
+        n_clusters = st.slider("Select Number of Clusters", 2, 10, 3, key="n_clusters")
+        kmeans = KMeans(n_clusters=n_clusters, random_state=42)
+        clusters = kmeans.fit_predict(word_freq_matrix)
+        
+        # Visualization
+        plot_data = pd.DataFrame({
+            'Folio': word_freq_matrix.index,
+            'Dim1': folio_coords[:, 0],
+            'Dim2': folio_coords[:, 1],
+            'Cluster': clusters
+        })
+        
+        fig = px.scatter(plot_data, x='Dim1', y='Dim2', color='Cluster', 
+                         hover_name='Folio', title=f"Folio Clustering ({dim_reduction_method})")
+        st.plotly_chart(fig)